The question of global inequality is perhaps the linchpin of all IPE debates for it concerns, among other things, the benefits of economic globalization and the efficacy of development efforts. For obvious reasons, those leaning left tend to say that global inequality is increasing, whereas those leaning right say the opposite. Among work representative of the former viewpoint is Thomas Pogge and Sanjay Reddy's, while that of the latter is Xavier Sala-i-Martin's. (Oddly enough, all three are at Columbia University; Pogge in the political science department, and Reddy and Sala-i-Martin in the economics department.) For this post, I will briefly describe the work of Branko Milanovic, which I consider more neutral. Even so, he gains a measure of respect from both sides of the debate. His book, Worlds Apart, is my reference on the matter.
Milanovic starts by tackling the question of what sort of inequality we are measuring. Concept 1 inequality is among the mean incomes of individual countries; tiny Lithuania population-wise counts for the same as large Russia. Concept 2 inequality weighs inequality according to each country's population size; Lithuania's average income would be extended to its 3.5M citizens, while Russia's average income to its 143M citizens. In an ideal world, we would have enough data to measure Concept 3 inequality, wherein we have data on the income of each individual in existence. While such data is nearly available in Western countries through household surveys, it is sparse in developing countries, to say the least.
The most commonly used measure of income inequality is the Gini index, which ranges from 0 (perfect equality) to 1 (perfect inequality where a single person has all the income.) Three main considerations also need to be accounted for, namely:
Let us move to Concept 1 inequality worldwide. In the graph above from p. 39 of his book, Milanovic uses GDP per capita in 1995 dollars and on PPP terms for 120 countries from 1950 to 2000. It is an open-and-shut case here: inequality among countries unweighted for population size is increasing. From 1982 onwards, Concept 1 inequality has been on the rise with poor countries doing worse on the average than rich ones. For what it's worth, this period coincides with the second oil shock caused by the fall of the Shah of Iran and rising interest rates worldwide as the US Federal Reserve took greater measures to curb inflation in America. Also, note the unfortunate boost given by African countries to the Gini index. Without them, the coefficient would be rather lower. Elsewhere, Milanovic highlights the contribution made to inequality by former Soviet-bloc countries for many of them became "downwardly mobile" in the wake of the Berlin Wall's fall.
The picture changes a lot when we consider the population weights of each country as in Concept 2 inequality using the same set of data previously described. The graph above from p. 87 depicts an improvement in the world Gini index, particularly over the last twenty years. Improvements in the economic performance of China and India mean that whereas they used to contribute to global income inequality, they now reduce it. As these two countries together account for over a third of the world's population with 1.3 and 1.1 billion persons respectively, their recent economic successes have made a large difference in these computations. However, critics note that China's GDP per capita may be inaccurate. To begin with, its GDP data is considered unreliable by many.
Given that China and India are such large countries, Milanovic considers whether Concept 2 results would differ if Chinese provinces and Indian states were substituted for simply "China" and "India" in the sample. Doing so makes sense. Consider what would happen if we used just one mean worldwide income to calculate the Gini index--there would be no inequality whatsoever. Segmenting the world population into finer groups would, all things equal, make measuring inequality a more accurate enterprise. That is, we would be moving towards Concept 3 and away from the more basic Concept 1. After doing so, Milanovic finds that "growing interregional inequality in China and India has a discernible and positive effect on world inequality," and that "as more Chinese (and Indian) provinces become rich while others stay behind, world inequality will rise" (p.99-100) . The Gini index is boosted by over five percentage points between 1980 and 2000 when Chinese provinces and Indian states are included in the sample.
Much, much more is to be found in Milanovic's masterful work. I cannot list more unless I intend to violate stipulations on fair use so I will end here. What Milanovic's work demonstrates to me at least is that inequality is a slippery concept that is highly sensitive to how you measure it. My ultimate take is that while China and India may have reduced Concept 2 inequality somewhat in recent years, this trend might be on the upswing again if regional inequalities in these two countries continue to grow. Sometimes the truth hurts, but it needs to be told.
Milanovic starts by tackling the question of what sort of inequality we are measuring. Concept 1 inequality is among the mean incomes of individual countries; tiny Lithuania population-wise counts for the same as large Russia. Concept 2 inequality weighs inequality according to each country's population size; Lithuania's average income would be extended to its 3.5M citizens, while Russia's average income to its 143M citizens. In an ideal world, we would have enough data to measure Concept 3 inequality, wherein we have data on the income of each individual in existence. While such data is nearly available in Western countries through household surveys, it is sparse in developing countries, to say the least.
The most commonly used measure of income inequality is the Gini index, which ranges from 0 (perfect equality) to 1 (perfect inequality where a single person has all the income.) Three main considerations also need to be accounted for, namely:
- Do we measure income at market exchange rates (usually against the US dollar) or on a purchasing power parity basis (PPP--what can actually be purchased locally)?
- Do we use survey-based mean income (from household surveys) or GDP per capita (which is basically GDP per head)?
- Do we measure income (what one earns) or expenditures (what one spends)? While Western and Latin American nations typically use income as an indicator, those in Africa and Asia typically use expenditures. The problem is that expenditures are fairly stable over time, presumably because basic needs have to be met, while income is more variable.
Let us move to Concept 1 inequality worldwide. In the graph above from p. 39 of his book, Milanovic uses GDP per capita in 1995 dollars and on PPP terms for 120 countries from 1950 to 2000. It is an open-and-shut case here: inequality among countries unweighted for population size is increasing. From 1982 onwards, Concept 1 inequality has been on the rise with poor countries doing worse on the average than rich ones. For what it's worth, this period coincides with the second oil shock caused by the fall of the Shah of Iran and rising interest rates worldwide as the US Federal Reserve took greater measures to curb inflation in America. Also, note the unfortunate boost given by African countries to the Gini index. Without them, the coefficient would be rather lower. Elsewhere, Milanovic highlights the contribution made to inequality by former Soviet-bloc countries for many of them became "downwardly mobile" in the wake of the Berlin Wall's fall.
The picture changes a lot when we consider the population weights of each country as in Concept 2 inequality using the same set of data previously described. The graph above from p. 87 depicts an improvement in the world Gini index, particularly over the last twenty years. Improvements in the economic performance of China and India mean that whereas they used to contribute to global income inequality, they now reduce it. As these two countries together account for over a third of the world's population with 1.3 and 1.1 billion persons respectively, their recent economic successes have made a large difference in these computations. However, critics note that China's GDP per capita may be inaccurate. To begin with, its GDP data is considered unreliable by many.Given that China and India are such large countries, Milanovic considers whether Concept 2 results would differ if Chinese provinces and Indian states were substituted for simply "China" and "India" in the sample. Doing so makes sense. Consider what would happen if we used just one mean worldwide income to calculate the Gini index--there would be no inequality whatsoever. Segmenting the world population into finer groups would, all things equal, make measuring inequality a more accurate enterprise. That is, we would be moving towards Concept 3 and away from the more basic Concept 1. After doing so, Milanovic finds that "growing interregional inequality in China and India has a discernible and positive effect on world inequality," and that "as more Chinese (and Indian) provinces become rich while others stay behind, world inequality will rise" (p.99-100) . The Gini index is boosted by over five percentage points between 1980 and 2000 when Chinese provinces and Indian states are included in the sample.
Much, much more is to be found in Milanovic's masterful work. I cannot list more unless I intend to violate stipulations on fair use so I will end here. What Milanovic's work demonstrates to me at least is that inequality is a slippery concept that is highly sensitive to how you measure it. My ultimate take is that while China and India may have reduced Concept 2 inequality somewhat in recent years, this trend might be on the upswing again if regional inequalities in these two countries continue to grow. Sometimes the truth hurts, but it needs to be told.
